Thermal Runaway / Fire / Newbie fears and questions...please take it easy on me!

[cyberfed]

Hello bright minds,

I have a fundamental question that I have been to embarrassed to ask, and it has me worried every since I asked in other post about paralleling batteries and I was shown a thread where a guy's house burnt down. Like all of us I don't want that to happen! I build packs such as these: 16S LEV60 LFP in a custome box with JBD BMS.
Since I am new at building packs this was a nice easy way for me to ease into building packs, and there's a detailed video on youtube on how to put it all together. I'm working on my 5th and final box of this type, then its on to my own customer pack builds with no templates of any kind.

I am scared of and have questions about thermal runaway and well...fire!

1. In a dead short or internal short of the battery, what happens in a series connection, the battery vents yes? Is that it if its able to vent? This is for just one battery and we are assuming the BMS and breaker on the box fail/dont work. Does the pack just become 15S? What exactly happens or can happen? (I know this is very situational but please give me some examples.

2. Dead short/internal short due to the protective covering of the batteries tearing and touching --I know this is more serious now we have 2 batteries that are in a scary condition. Again lets assume the BMS/breaker fail or dont work. What happens? Thermal runaway and fire at that point?

3. Now lets assume the BMS and breaker work -- what impact (if any) would they have on #1 and #2 scenarios? In particular to the battery pack.

4. If you take a look at that link and poke around a bit, you will see that it uses a PCB for all the connections. Meaning after I put the batteries in and tighten the compression plates the only wires I have to rig up are main postitive and main negative from the PCB to the the BMS which goes out to the terminals outside the box. I'm concerned in a internal short situation the PCB would just melt and become one giant conductor and just bad news bears. Unfortunately the company who sells this stuff while they are great people dont sit there and provide technical support such as this, it is very much sold as a DIY disclaimer. They have a FB group where I guess you can ask questions but I don't have an account and don't really wanna make one, I will if I have to.

I wanted to put FR-4 boards in between the batteries as extra protection but the tolerances of the box/compression plate don't allow it, they are too think (I bought .4 mm plates). Then I was going to use fish tape and measured that at .2mm so that wouldn't work either. I have been left using Kapton tape to add another layer of protection between the batteries, I simply can't make them fit with anything more. I did double up the kapton on all 8 edges of the batteries as I've heard that is the weakspot.

The packs themselves dont move around much in the house, they have their own metal cart with wheels to be moved around the house for charging or powering something ect.. batteries are brand new, the IR was almost identical on all of them, and they balance during charge/discharge to 3ma. There was a lot of talk of T-fuses in the other post but that was due to the fact I was asking about parallel batteries. In this case since its just one big series pack, what does a T-fuse get me if there's a problem with the batteries themselves? Ok the fuse blows, but the batteries physical situation hasn't changed, with a blow fuse it just can't pump that out the pack. If need be I'm willing to drill a hole on the top of the box to add a t-fuse to the top of the box, but just dont see what I get outside of maybe protection of downstream items which is pointless if the 16S pack is going burst into flames.

Would love to hear opinions on this. Is the steel enclosure enough? I mean this not super thick and if a fire broke out it would be HOT and I'm sure melt the metal. I've been thinking of adding some temperature probes to the outside of the box in various places and measure the temp, if it gets higher than a certain level, set off an alarm I can hear so I can react. Packs are only used when someone is home and never have even a remotely high load ( less than 100w ). So I can't imagine its stressing the batteries hard. The BMS unfortunately has no BT or wireless monitoring capability. I'm good with microcontrollers so I can rig some up with temp sensors to essentially act as a fire/ heat sensor to warn me. I purchased a 55lb bag of CellBlockEX, if you are not familiar with it, its designed to put out Lithium Ion fires. I figured between extengishers and that stuff I could dump on the pack if god forbid something when south it should stop things.

Like I said I have the T-fuses on hand, cost me a fortune. I also bought one (just to test it out) terminal mount style fuse (I know its nowhere near as good as t-fuses) from Blue Sea Systems. It actually mounts very nicely to the outside of that enclosure. So with that installed I have the BMS, a cheap breaker that is part of the kit I dont trust, and a fuse on the terminal. In my mind none of this stops the batteries themselves if things go south, maybe if only one goes bad. I dunno am I over thinking this. This fire thread that was shared got in my head. I'm not doing parallel batteries. Please take a look at the link I shared (safe I promise) so you can see the box, PCB and batteries ect.. that are used in the build so you have an idea with with I'm working with. If I need to enclose the whole box in firebricks or something else I'll do it. Like I said I'm on my 5th and final one of these builds, working on the kapton tape wrap of the batteries now. Then its done and then its on to more conventional style builds (using wires between each battery ect..).


Thanks in advance folks
Cyberfed

 

[robbob2112]

Hello bright minds,

I have a fundamental question that I have been to embarrassed to ask, and it has me worried every since I asked in other post about paralleling batteries and I was shown a thread where a guy's house burnt down. Like all of us I don't want that to happen! I build packs such as these: 16S LEV60 LFP in a custome box with JBD BMS.
Since I am new at building packs this was a nice easy way for me to ease into building packs, and there's a detailed video on youtube on how to put it all together. I'm working on my 5th and final box of this type, then its on to my own customer pack builds with no templates of any kind.

I am scared of and have questions about thermal runaway and well...fire!

1. In a dead short or internal short of the battery, what happens in a series connection, the battery vents yes? Is that it if its able to vent? This is for just one battery and we are assuming the BMS and breaker on the box fail/dont work. Does the pack just become 15S? What exactly happens or can happen? (I know this is very situational but please give me some examples.

If a single cell goes into TR (thermal runaway) and vents it releases a bunch of different gassess but 50% of that roughly is hydrogen. If it gets a spark or touches something that is 300c hot it will catch fire. The cells next to the TR cell may also vent and add to the problem. If the cause was overcharging/overvoltage this is likely since all the cells will be very close to this point. If the cause was physical damage to the one pack it is a coin toss.

As the resistance in the pack drops the other parallel packs would dump current into the bad pack.

2. Dead short/internal short due to the protective covering of the batteries tearing and touching --I know this is more serious now we have 2 batteries that are in a scary condition. Again lets assume the BMS/breaker fail or dont work. What happens? Thermal runaway and fire at that point?

A short between two cells casings when they are wired in a typical configuration is the same as laying a screwdriver across the pos and neg - the cell will short and put all its current out as quick as possible which will cause a vent - see above

3. Now lets assume the BMS and breaker work -- what impact (if any) would they have on #1 and #2 scenarios? In particular to the battery pack.

The BMS or breaker will stop other packs from shorting into the bad one (lower resistance) and hopefully there isn't a spark after the vent - see above.

4. If you take a look at that link and poke around a bit, you will see that it uses a PCB for all the connections. Meaning after I put the batteries in and tighten the compression plates the only wires I have to rig up are main postitive and main negative from the PCB to the the BMS which goes out to the terminals outside the box. I'm concerned in a internal short situation the PCB would just melt and become one giant conductor and just bad news bears. Unfortunately the company who sells this stuff while they are great people dont sit there and provide technical support such as this, it is very much sold as a DIY disclaimer. They have a FB group where I guess you can ask questions but I don't have an account and don't really wanna make one, I will if I have to.

The PCB material used is typically flame resistant - it will burn if you have a blow torch on it, but when you take the flame away it will stop. What will most likely happen if the BMS/fuse don't stop the current it will melt the runs on the board. See above for vent.

I wanted to put FR-4 boards in between the batteries as extra protection but the tolerances of the box/compression plate don't allow it, they are too think (I bought .4 mm plates). Then I was going to use fish tape and measured that at .2mm so that wouldn't work either. I have been left using Kapton tape to add another layer of protection between the batteries, I simply can't make them fit with anything more. I did double up the kapton on all 8 edges of the batteries as I've heard that is the weakspot.

The packs themselves dont move around much in the house, they have their own metal cart with wheels to be moved around the house for charging or powering something ect.. batteries are brand new, the IR was almost identical on all of them, and they balance during charge/discharge to 3ma. There was a lot of talk of T-fuses in the other post but that was due to the fact I was asking about parallel batteries. In this case since its just one big series pack, what does a T-fuse get me if there's a problem with the batteries themselves? Ok the fuse blows, but the batteries physical situation hasn't changed, with a blow fuse it just can't pump that out the pack. If need be I'm willing to drill a hole on the top of the box to add a t-fuse to the top of the box, but just dont see what I get outside of maybe protection of downstream items which is pointless if the 16S pack is going burst into flames.

I would still use a class T fuse personally. When a class T fuse burns the sand inside fills in the gap and extenguishes the arc. This is assuming you had something like an external short it would provide protection in that case.

Would love to hear opinions on this. Is the steel enclosure enough? I mean this not super thick and if a fire broke out it would be HOT and I'm sure melt the metal. I've been thinking of adding some temperature probes to the outside of the box in various places and measure the temp, if it gets higher than a certain level, set off an alarm I can hear so I can react. Packs are only used when someone is home and never have even a remotely high load ( less than 100w ). So I can't imagine its stressing the batteries hard. The BMS unfortunately has no BT or wireless monitoring capability. I'm good with microcontrollers so I can rig some up with temp sensors to essentially act as a fire/ heat sensor to warn me. I purchased a 55lb bag of CellBlockEX, if you are not familiar with it, its designed to put out Lithium Ion fires. I figured between extengishers and that stuff I could dump on the pack if god forbid something when south it should stop things.

No opinion the cellblockex -

Like I said I have the T-fuses on hand, cost me a fortune. I also bought one (just to test it out) terminal mount style fuse (I know its nowhere near as good as t-fuses) from Blue Sea Systems. It actually mounts very nicely to the outside of that enclosure. So with that installed I have the BMS, a cheap breaker that is part of the kit I dont trust, and a fuse on the terminal. In my mind none of this stops the batteries themselves if things go south, maybe if only one goes bad. I dunno am I over thinking this. This fire thread that was shared got in my head. I'm not doing parallel batteries. Please take a look at the link I shared (safe I promise) so you can see the box, PCB and batteries ect.. that are used in the build so you have an idea with with I'm working with. If I need to enclose the whole box in firebricks or something else I'll do it. Like I said I'm on my 5th and final one of these builds, working on the kapton tape wrap of the batteries now. Then its done and then its on to more conventional style builds (using wires between each battery ect..).

You are right, if mistakes are made it can be dangerous - check everything - this amount of stored energy has a lot of potential.

Thanks in advance folks
Cyberfed

 

[k490]

I bought 11G steel from local steel supply when my battery box is done it will weigh almost 300 lbs before putting the batteries in. Also have a 4" vent pipe goes out through a solid block wall. If it didn't get so hot here in Phoenix I would put the batteries outside in a steel box. I was going to go with 1/4" thick plate steel for the top but lifting 120lb lid off the box wasn't really feasible. I will still use class T fuses and other safety measures in case all that fails the box will be able to keep any fire from escaping while venting any explosive gas outside.

 

[cyberfed]

Thanks folks, I appreciate the feedback, keep it coming if you have other ideas/opinions! Just a reminder because it sounded like @robbob2112 thought I was paralleling the packs. I am not. They are single packs in the metal enclosure 16S LFP. Re-reading what he wrote I think he meant each battery cell...I think.

Very interesting about the PCB theory, I though it would be a hinderance vs helpful. I purchased some ceramic fiberglass that is fire resistant to 2000F its also non conductive. I was thinking of stuffing that inside in the voids of the case, not that there's much voids its pretty packed once built, I wish I could fit a class T fuse in there but there's just no room. Or is that a bad idea (the fiberglass)?

I guess the best plan would be is to 1. drill a small hole big enough for the wires from the battery POS to run out of the box, through a class T fuse I already have and then back down into the box where it connects inside the case. That way its a direct fused connection at the battery POS.

2. Also drilling another hole to mount a small fan that sucks air out (with all the dog hair I certainly dont want to go the opposite way and blow air in) so if there is a vent at least we are clearing out the combustible gases. I can even put a sensor by that exhaust that will trigger an alarm if it detects the gases (I've actually already built one with a microcontroller for fun its somewhere around the house lol). Wherever I have a pack I have a healthy amount of CellBlockEX as well as an extinguisher at the ready. Bless my girlfriends heart for allowing me to turn out house into batteries all over lol.

@k490 -- I have had the same thoughs of reinforcing the box with thicker steel or some other method. I guess I need to find a steel place around me and talk to them about how I could get what I need as I don't have tools or the knowledge to properly cut/bend metal. I keep all the cases on rolling moving metal carts so weight isn't an issue. They are rated for like 1200lbs or something. I too, like yourself, would like to keep stuff outside but in Florida with the heat and rain its just not going to work. Especially since I rent.

I don't know the cost of custom steel like that or what I would need,3. I guess all sides.. I looked online once at some custome metal place and it was crazy expensive, not to mention you would be surprised at the melting point of most steels. Its low, well I'm thinking compared to what a battery fire would create, which to be fair I don't know what that temp range is. ----NM just saw above 300C, that's actually WAY lower than I thought that is managable to contain. If you have an arcing spot that has to be well over the melting point of steel unless its seriously thick. These particular batteries can dump 600A so there's some serious energy that could be dumped.

Sounds like the key is exhaust, protection (metal/something) and fusing.

I've created a ghetto version of protection on my ANL fuses outside the box that go to the inverter and SCC. It's essentially me trying to recreate a class-T type fuse. (until I can get enough class T's to cover everything). Let me know if this sounds like a bad idea, or could potentially help in the dreaded arcing of a toasted fuse such as an ANL.

I took a fair amount of CellBlockEX (its like a sand/glass pellets mix type stuff made for lithium ion fire extinguishment) in the top of the plastic ANL fuse lid. Then I added some of the ceramic 2000F (I've tested this with a blow torch and thermal camera, it does very well keeps the heat isolated and then slowly disperses it outward to the rest of the insulation, cools pretty quickly, and didn't catch fire just got a little black) fiberglass on top of the CellBlockEX pellets as a barrier to hold them from falling all over the place. I then used kapton tape to hold the edges around the cable entry points to again keep everything from falling out, packed it tight with the insulation and its closes very snuggly with the fuse installed/wired. Then I did a wrap of kapton tape around the outside of the plastic box to hold it from accidentally opening and spilling everything.

My thinking is this: Something goes wrong with the inverter or the load, whatever and we end up in a situation where the fuse blows and is causing an arc condition, or will cause an arc condition. My ghetto ANL fused protection would essentially fall on top of the blown fuse and the insulation likely would prevent the arc from even starting, but lets say it does arc and burns through the insulation then the layer of CellblockEX pellets would drop on it, essentially like dropping sand on it. In hopes it would stop the arc. Like I said this is ghetto I know, but does it likely increase or decrease the chances of an arc in an ANL fuse? I know its just a what it scenario.

Again this is just a cheapo band aid until I get get T-fuses everywhere. As it stands right now between the batteries to the inverter (taking the SCC out of the equation because that is used just for charging and then removed from the circuit but it follows the same level of safety measures) I have Battery--->BMS--->DC Breaker (Cheap Chinese junk I wouldn't trust, cant find a quality one that fits the spot for it on the box)---->Blue Sea Systems Terminal fuse at the main POS of the battery box-->disconnect switch--->followed by the ANL fuse near the inverter--->inverter--->load. Inverters are Victron so they aren't total junk I know people have strong feelings about them both ways. The portion of it that touches the metal has been coated with a non conductive chemical (its from MG I forget the name its weatherproof/nonconductive and it works I've stress tested it). Nearby by is a bunch of CellblockEX to pour on anything if need be, and fire extinguishers.

I'm also going to go into the JBD BMS and change the short circuit value and lower it down, its at 1400A right now, I want bring that way down so it triggers at a much lower amperage.

Packs are never overcharged or discharged, the BMS actually does a very good job and stopping that from happening and the Victron inverters also allow you to set a shutoff voltage (tested and works). So two layers to keep the batteries from dropping too low. I only charge stuff when I'm home to keep an eye on it and its just a 4x 100 watt panel series setup to give me the 48V to charge the packs, so the max amperage I can get on a perfect sunny day with perfect everything is 5.9 amps, usually its around 4.5-5 since the sun is moving around and I'm not constantly re-adjusting the angles. Lastly the loads the inverters are connected to are super low, 50 watts maybe 70 watts, they just power small things like our lights and things like that. I'm not sure I've ever gone over 1 amp being pulled from the battery when manually checking. So really my biggest concern is internal battery failure of some sorts. They are brand new batteries that were meant for BMW hybrids so they are high quality Japanese batteries and I tested them prior to install, they have all been great, the IR (measured with a real tool not some cheap battery charger) is consistent across the batteries and same with the voltage when they arrived. That said I know anything can happen so I'm not letting my guard down.

As always appreciate any feedback, good or bad, you may have, the goal is safety!
Cheers,
Cyberfed

 

[Ampster]

A short between two cells casings when they are wired in a typical configuration is the same as laying a screwdriver across the pos and neg

As a precaution I used thin plastic file folders as dividers. My pack was 3P16S so I only did every third cell because there should be no potential betwee cells in parallel. One member did measure the current between cell casings of cells in series and reported that the current was small and there was little or no continuity between the casing and either the positive or the negative terminals so I doubt that the current between casings would be the same as a dead short but it could be enough to unbalance the pack.

 

[cyberfed]

Thank you @Ampster! Do you think kapton tape is good enough to add to the factory wrap? I double tape the 8 corners so its multi layered. Curious, have you measured your plastic file folders with calipers and what size are they in mm? I dont have almost any room to work with.

I guess my reasoning for adding something to begin with aside that its a best practice, is that I'm afraid of the original wrap breaking on multiple batteries creating a short. That is the suspected theory of how this one thread lost his home to a fire after his system worked perfectly for over 4 years is that the wraps likely tore and touched creating a short and from there total destruction, there's more to it than that but no one knows the true cause not even the fire inspector. So I'm paranoid.

I'm using compression plates in the box but the tolerances are so tight I can't fix FR boards in between each battery it leaves me well over a quarter of an inch too thick and thats of course only placing one piece between 2 batteries not doubling up. I did the math on the fishtape and that would be too thick as well. So Kapton is what I've gone with. It's physically impossible to make more room the way the metal box is made. Just hoping to add another layer of safety to the orginal wrap. It seems to be the same type of wrap that many of my 18650 cells have and I've abused some of those hard and they hold up well. My packs of course are sealed shut in the box and aren't being abused. But I worry over time if that wrap would deteriorate. Regular visual inspections will be done as well.

 

[Ampster]

That is the suspected theory of how this one thread lost his home to a fire after his system worked perfectly for over 4 years is that the wraps likely tore and touched creating a short and from there total destruction, there's more to it than that but no one knows the true cause not even the fire inspector.

I did not read all 50 pages of that thread but if that theory was offered, it is inconsistant with the other members observation that the current is rather small between the case and there was no continuity between either positive or negative terminal and the case. I do agree it is best practice and the reason I used something is to avoid a small discharge which could possible create an imbalance. I did not measure my folders but they were about the same thickness as the black plastic on the top or the cells. It was a lot easier to cut fifteen of them than tape even one side of every cell.

 

[SparkyJJO]

If there is no continuity between either cell positive and case or cell negative and case, how can there be a short circuit between cases?

 

[upnorthandpersonal]

I'm still not convinced that the cell case of certain manufactures would be directly connected to either positive or negative cell terminal. I have observed some capacitive coupling, but nothing that would indicate you can short anything out. I have tested this with EVE280 cells in the past; I'll try with some others if I find time.

 

[k490]

I got all the steel shear cut by the steel supply nothing to cut they are able to make perfect cuts with the shear cutting machine all I need to do is weld it together. Since I already have a Mig welder not a problem for me the total cost was $216 including all the cuts and steel. You can hire a mobile welder come and weld it together for you. You could line the inside with fireboard also.